BIOL281 Lecture 1: Biochem Tests PDF

BIOL281 Lecture 1: Use of Biochemical Tests in Medicine It’s All About Quality! Dr Carole Rolph [email protected] Dr. Christine Shirras Office: B76 Furness Email: [email protected] Learning Objectives Ident...

BIOL281 Lecture 1: Use of Biochemical Tests in Medicine It’s All About Quality! Dr Carole Rolph [email protected] Dr. Christine Shirras Office: B76 Furness Email: [email protected] Learning Objectives Identify appropriate sample tubes for blood where a glucose test, whole blood analysis, or general serum biochemical tests will subsequently be performed. Define the term ‘reference range’ or ‘reference interval’ and explain what this means in practice, regarding what percentage of results will be within 1 SD, 2SD and 3SD of the mean. Define: analytical sensitivity, analytical specificity, accuracy, precision, Limits of linearity, coefficient of variation, action limit, random error, systematic error (bias). Learning Objectives Given data output from two assays, Draw an Altman-Bland plot, explain the purpose of this plot and interpret your plot. Draw a precision profile for two assays and explain why this would be used. Given a set of ‘blank’ data, explain how a detection limit is set. Explain the purpose of a Levy-Jennings plot and interpret data presented in this way. Explain the process, purpose and importance of both internal and external quality control. Reading and see MyResource List Ahmed, N. Fundamentals of Biomedical Science Clinical Biochemistry. (2011 or 2017) Chapter 1. pp1-30 (e-book) Gaw. Clinical Biochemistry: an illustrated colour text. (2013) Pp 4-10, (e-book) Quality, Quality, Quality…… Performance, Performance, Performance Performance of assay Performance of different analytical methods Day to day performance of equipment for a specific assay Internal Quality Control External Quality Control Role of the clinical biochemistry lab Provision of results which are: Timely Accurate Reproducible Relevant Expert interpretative advice Acceptance or rejection of results What is the purpose of clinical biochemistry tests? Diagnosis – e.g. myocardial infarction Prognosis - predict clinical outcome of disease Screening - identification of undiagnosed clinical condition e.g. newborn screening phenylketonuria (PKU) Monitoring Determining treatment – e.g. diabetes mellitus Monitoring treatment – e.g. thyroid hormone replacement Monitoring long term/chronic condition – e.g. PSA Typical Sample Workflow Diagram Types of Samples for Biochemical Testing Blood (serum, plasma) Urine Faeces CSF Saliva Sweat Hair Sample collection and delivery to the lab Results for a sample must be meaningful Appropriate sample must be sent to the pathology laboratory: blood, urine, CSF, faeces, saliva, Correct tubes/bottles must be used for sample collection Appropriate timing of samples e.g. Fasting blood glucose or random? Request form properly completed Delivered quickly to the lab. Why? Tubes used for blood collection SST gel Blood Glucose Analysis Potassium Oxalate Anticoagulant Binds calcium ions Sodium Fluoride Inhibits glycolysis Maintains the original glucose and lactate concentrations Diabetes and It’s Outcomes World Health Organisation Report published in 2016 Data Thecollected in 2014. starting point for diagnosis is a blood glucose assay. Monitoring of blood glucose is a vital part of diabetes management. https://i.dailymail.co.uk/i/pix/2016/04/06/13/32E6EE6300000578-3526296- Reference interval : The_number_of_adults_with_diabetes_has_quadrupled_to_422_million-m- 70_1459946310792.jpg Reference intervals/ranges Reference intervals chosen to include 95% of the values found in healthy volunteers. So… 5% of the normal population will be outside the reference range! 67% - mean ± 1 SD 95% - mean ± 2 SD 99% - mean ± 3 SD Reference interval = mean ± 2 SD Each lab has its own reference interval Spectrophotometric ways of measuring blood glucose concentration Assay 1 Assay 2 Direct chemical Enzyme linked reaction reaction React glucose with a Glucose oxidase oxidises chemical to get a glucose and releases hydrogen peroxide colour change Hydrogen peroxide then e.g reacts with o-Danisidine dinitrosalicylic acid (D to give a coloured NS) product  For each assay we must know: Analytical Sensitivity The smallest amount of the analyte which can be detected (detection limit) Analytical Specificity How well an assay detects only a specific substance and does not detect closely related substances. Accuracy How close the result comes to the true value. Precision Reproducibility of multiple measurements. Limits of Linearity a range of concentrations between which the result can be reported as accurate. Precision and accuracy Precision = reproducibility Accuracy = how close to the measured value is the actual value? Ahmed p 24 (2011) or 23 (2017) Imprecise Precise but Precise and accurate (Random error)) inaccurate (systematic error, bias) Regression Analysis: Comparison of the output of two different assays, measuring a given analyte in the same samples Assay 2 Glucose conc (mmoles/L) Assay 1 Glucose conc (mmoles/L ) Graph indicates a relationship between the data produced from the two assays, but does not indicate how closely the assays are in agreement. Bland – Altman Plot (aka Difference plot) : Indicates whether there is agreement between two assays. + 2SD Assay 1 – Assay 2 value - 2SD Mean (assay 1 + 2) / 2 Typical results from the 2 assays Bland-Altman Worksheet Bland-Altman Plot  Determination of a Precision profile across two assays Concentration of analyte in this example = 1 mmole/L Assay 1 output Assay 2 output 0.0020 0.0018 CV = Coefficient of Variation 0.0015 0.0020 0.0026 0.0017 CV= (SD/mean) x 100 0.0011 0.0019 0.0009 0.0019 Assay 1 CV = 32.5% 0.0017 0.0018 0.0016 0.0020 Assay 2 CV = 6.3% 0.0010 0.0021 0.0020 0.0020 0.0018 0.0019 Average = 0.0016 Average = 0.0019 SD = 5.2 e-4 SD = 1.2 e -4 Repeatat other concentrations e.g. 2 mmole/L and 3 mmole/L Precision profile for the two different assays, measuring the same analyte at three different concentrations. What can you say about the precision of these assays? 35 30 25 20 CV (%) 15 Assay 1 10 Assay 2 5 0 0.5 1 1.5 2 2.5 3 3.5 Glucose concentration mmole/L  Detection limit (sensitivity) Lowest amount of analyte which can be distinguished from absence of analyte. Analyse a mock sample with none of analyte present. Output readings when no analyte present might be: 0.012 0.028 Mean = 0.0197 0.018 SD = 0.0051 0.022 Mean + 2 SD= 0.0299 0.014 0.025 Estimate amount of analyte 0.022 which would give this reading. 0.018 0.015 This is the detection limit. 0.023  Detection Limits Signal Detection limit Concentration  Internal Quality Control Are the results today the same as yesterday? Quality control materials manufactured. Analyte in question is provided at various concentrations. H, M,L QC materials tested on machine frequently each day and after changing lamp, reagent etc. Data plotted on a Levy-Jennings Plot Levy-Jennings chart indicates performance of a test with quality control samples, over time Glucose conc (mmoles/L) 8 + 2SD 6 Mean 4 2 - 2SD 0 Days Results for a Glucose 5 mmol/L QC sample  Levy-Jennings Plot -Trend and Shift d = 6 or more points show a consecutive move in the same tion = 6 or more results are on one side of the mean rather than ered about the mean. (i.e. The accuracy of the test has ged – mean of the values is not the true value). Shift P 24 Ahmed 2017 P26 Ahmed 2011 Mean Trend Targeting of errors Random error Accurate but imprecise (mean value = accurate) Systematic error (Positive bias) Precise but inaccurate Accurate and precise P23 Ahmed 2017 P24 Ahmed 2011 External Quality Control Are my results the same as those of other labs performing the same test? e.g UKNEQAS (UK National External Quality Assessment Service) and others. Quality control samples regularly sent out to all labs. Performance reports generated Assesses performance of a lab Compares performance of lab with other labs using the same analyser and method Compare against other analysers and methods measuring the same analyte. Precision? Bias? Example NEQAS Report Point of Care Testing Defined as : Laboratory tests that are performed at or near a patient at or near a patient and at a site where care or treatment may be provided. Revolutionised the way in which patients with diabetes are able to manage their condition. How can QC be managed with instruments bought on the internet ? Patient training Interpretation of results Is the instrument accurate? Is the instrument precise? Servicing of the instrument?